Company Description
Eytrix is a cutting-edge technology company that enables governments and enterprises to leverage AI for security camera systems. Our seamless interface allows users to query footage, set up intelligent alerts, and extract data from video streams using natural language commands.
Role Description
This is a full-time ML Engineer - AI/Computer Vision role at Eytrix located in Los Angeles, CA. The ML Engineer will be responsible for developing and implementing AI algorithms for computer vision, pattern recognition, and neural networks. The role involves working on-site to enhance the company's AI-powered interface and improve video analysis capabilities.
What You’ll Do
- Design end‑to‑end ML & LLM pipelines: Data ingestion, labeling, synthetic dataset generation, model registry, and rollout.
- Build multimodal VLM services: enable natural‑language querying, scene description, and system reasoning.
- Optimize real‑time video analytics: Profile GPU workloads, quantize/compile models with TensorRT, and stream inference results at sub‑second latency.
- Drive MLOps best practices: Continuous training, drift detection, self‑healing rollbacks, and A/B testing of models in production.
- Collaborate cross‑functionally: Partner with CV researchers, LLM prompt engineers, and backend teams to expose model insights via gRPC/REST and event streams.
- Advance safety & compliance: Implement zero‑trust networking, secret management, and audit trails for our multi‑tenant clusters in high‑security environments.
What We’re Looking For
- 3 + years building ML or MLOps systems.
- Production experience with Docker & Kubernetes and modern CI/CD.
- Strong backgorund in computer‑vision pipelines (classification, detection, tracking) and common frameworks (PyTorch, TensorFlow, OpenCV).
- Comfort profiling GPU workloads and squeezing every millisecond from CUDA/TensorRT or similar.
Big Bonuses
- Proven track record deploying and maintaining on‑prem or edge solutions.
- Hands‑on experience fine‑tuning or integrating LLMs (RAG, LoRA, agents).
- Knowledge of stream‑processing for high‑volume video data.
- Background working with government or regulated customers where security and auditability are strict.
- Experience Rust or Go
